Hands-on Exercise 3: Processing and Visualising Flow Data

Overview

Spatial interaction represent the flow of people, material, or information between locations in geographical space. It encompasses everything from freight shipments, energy flows, and the global trade in rare antiquities, to flight schedules, rush hour woes, and pedestrian foot traffic.

Each spatial interaction, as an analogy for a set of movements, is composed of a discrete origin/destination pair. Each pair can be represented as a cell in a matrix where rows are related to the locations (centroids) of origin, while columns are related to locations (centroids) of destination. Such a matrix is commonly known as an origin/destination matrix, or a spatial interaction matrix.

In this hands-on exercise, you will learn how to build an OD matrix by using Passenger Volume by Origin Destination Bus Stops data set downloaded from LTA DataMall. By the end of this hands-on exercise, you will be able:

• to import and extract OD data for a selected time interval,

• to import and save geospatial data (i.e. bus stops and mpsz) into sf tibble data frame objects,

• to populate planning subzone code into bus stops sf tibble data frame,

• to construct desire lines geospatial data from the OD data, and

• to visualise passenger volume by origin and destination bus stops by using the desire lines data.

Getting Started

For the purpose of this exercise, these R packages will be used:

• sf for importing, integrating, processing and transforming geospatial data.

• tidyverse for importing, integrating, wrangling and visualising data.

• tmap for creating thematic maps.

pacman::p_load(tmap, sf, DT, stplanr,
               performance,
               ggpubr, tidyverse)

also installing the dependencies 'BH', 'nabor', 'od', 'pbapply'

Warning: unable to access index for repository http://www.stats.ox.ac.uk/pub/RWin/bin/windows/contrib/4.3:
  cannot open URL 'http://www.stats.ox.ac.uk/pub/RWin/bin/windows/contrib/4.3/PACKAGES'

package 'BH' successfully unpacked and MD5 sums checked
package 'nabor' successfully unpacked and MD5 sums checked
package 'od' successfully unpacked and MD5 sums checked
package 'pbapply' successfully unpacked and MD5 sums checked
package 'stplanr' successfully unpacked and MD5 sums checked

The downloaded binary packages are in
    C:\Users\peish\AppData\Local\Temp\RtmpwXxJDp\downloaded_packages

stplanr installed

Preparing the Flow Data

Importing the OD data

Firstly, we will import the Passenger Volume by Origin Destination Bus Stops data set downloaded from LTA DataMall by using read_csv() of readr package.

odbus <- read_csv("data/aspatial/origin_destination_bus_202310.csv")

Rows: 5694297 Columns: 7
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (5): YEAR_MONTH, DAY_TYPE, PT_TYPE, ORIGIN_PT_CODE, DESTINATION_PT_CODE
dbl (2): TIME_PER_HOUR, TOTAL_TRIPS

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Let’s display the odbus tibble data table by using the code chunk below.

glimpse(odbus)

Rows: 5,694,297
Columns: 7
$ YEAR_MONTH          <chr> "2023-10", "2023-10", "2023-10", "2023-10", "2023-…
$ DAY_TYPE            <chr> "WEEKENDS/HOLIDAY", "WEEKDAY", "WEEKENDS/HOLIDAY",…
$ TIME_PER_HOUR       <dbl> 16, 16, 14, 14, 17, 17, 17, 7, 14, 14, 10, 20, 20,…
$ PT_TYPE             <chr> "BUS", "BUS", "BUS", "BUS", "BUS", "BUS", "BUS", "…
$ ORIGIN_PT_CODE      <chr> "04168", "04168", "80119", "80119", "44069", "2028…
$ DESTINATION_PT_CODE <chr> "10051", "10051", "90079", "90079", "17229", "2014…
$ TOTAL_TRIPS         <dbl> 3, 5, 3, 5, 4, 1, 24, 2, 1, 7, 3, 2, 5, 1, 1, 1, 1…

A quick check of odbus tibble data frame shows that the values in ORIGIN_PT_CODE and DESTINATION_PT_CODE are in numeric data type. Hence, the code chunk below is used to convert these data values into character data type.

odbus$ORIGIN_PT_CODE <- as.factor(odbus$ORIGIN_PT_CODE)
odbus$DESTINATION_PT_CODE <- as.factor(odbus$DESTINATION_PT_CODE) 

Extracting the study data

For the purpose of this exercise, we will extract commuting flows on weekday and between 6 and 9 o’clock.

odbus6_9 <- odbus %>%
  filter(DAY_TYPE == "WEEKDAY") %>%
  filter(TIME_PER_HOUR >= 6 &
           TIME_PER_HOUR <= 9) %>%
  group_by(ORIGIN_PT_CODE,
           DESTINATION_PT_CODE) %>%
  summarise(TRIPS = sum(TOTAL_TRIPS))

`summarise()` has grouped output by 'ORIGIN_PT_CODE'. You can override using
the `.groups` argument.

Table below shows the content of odbus6_9

datatable(odbus6_9)

Warning in instance$preRenderHook(instance): It seems your data is too big for
client-side DataTables. You may consider server-side processing:
https://rstudio.github.io/DT/server.html

We will save the output in rds format for future use:

write_rds(odbus6_9, "data/rds/odbus6_9.rds")

The code chunk below will be used to import the saved odbus6_9.rds into R environment.

odbus6_9 <- read_rds("data/rds/odbus6_9.rds")

Working with Geospatial Data

For the purpose of this exercise, two geospatial data sets will be used. They are:

• BusStop: This data provides the location of bus stop as at last quarter of 2022.

• MPSZ-2019: This data provides the sub-zone boundary of URA Master Plan 2019.

Both data sets are in ESRI shapefile format.

Importing geospatial data

busstop <- st_read(dsn = "data/geospatial",
                   layer = "BusStop") %>%
  st_transform(crs = 3414)

Reading layer `BusStop' from data source 
  `C:\PeiShan0502\ISSS624\Hands-on_Ex3\data\geospatial' using driver `ESRI Shapefile'
Simple feature collection with 5161 features and 3 fields
Geometry type: POINT
Dimension:     XY
Bounding box:  xmin: 3970.122 ymin: 26482.1 xmax: 48284.56 ymax: 52983.82
Projected CRS: SVY21

mpsz <- st_read(dsn = "data/geospatial",
                   layer = "MPSZ-2019") %>%
  st_transform(crs = 3414)

Reading layer `MPSZ-2019' from data source 
  `C:\PeiShan0502\ISSS624\Hands-on_Ex3\data\geospatial' using driver `ESRI Shapefile'
Simple feature collection with 332 features and 6 fields
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: 103.6057 ymin: 1.158699 xmax: 104.0885 ymax: 1.470775
Geodetic CRS:  WGS 84

mpsz

Simple feature collection with 332 features and 6 fields
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: 2667.538 ymin: 15748.72 xmax: 56396.44 ymax: 50256.33
Projected CRS: SVY21 / Singapore TM
First 10 features:
                 SUBZONE_N SUBZONE_C       PLN_AREA_N PLN_AREA_C       REGION_N
1              MARINA EAST    MESZ01      MARINA EAST         ME CENTRAL REGION
2         INSTITUTION HILL    RVSZ05     RIVER VALLEY         RV CENTRAL REGION
3           ROBERTSON QUAY    SRSZ01  SINGAPORE RIVER         SR CENTRAL REGION
4  JURONG ISLAND AND BUKOM    WISZ01  WESTERN ISLANDS         WI    WEST REGION
5             FORT CANNING    MUSZ02           MUSEUM         MU CENTRAL REGION
6         MARINA EAST (MP)    MPSZ05    MARINE PARADE         MP CENTRAL REGION
7                   SUDONG    WISZ03  WESTERN ISLANDS         WI    WEST REGION
8                  SEMAKAU    WISZ02  WESTERN ISLANDS         WI    WEST REGION
9           SOUTHERN GROUP    SISZ02 SOUTHERN ISLANDS         SI CENTRAL REGION
10                 SENTOSA    SISZ01 SOUTHERN ISLANDS         SI CENTRAL REGION
   REGION_C                       geometry
1        CR MULTIPOLYGON (((33222.98 29...
2        CR MULTIPOLYGON (((28481.45 30...
3        CR MULTIPOLYGON (((28087.34 30...
4        WR MULTIPOLYGON (((14557.7 304...
5        CR MULTIPOLYGON (((29542.53 31...
6        CR MULTIPOLYGON (((35279.55 30...
7        WR MULTIPOLYGON (((15772.59 21...
8        WR MULTIPOLYGON (((19843.41 21...
9        CR MULTIPOLYGON (((30870.53 22...
10       CR MULTIPOLYGON (((26879.04 26...

• st_read() function of sf package is used to import the shapefile into R as sf data frame.

• st_transform() function of sf package is used to transform the projection to crs 3414.

The code chunk below will be used to write mpsz sf tibble data frame into an rds file for future use.

mpsz <- write_rds(mpsz, "data/rds/mpsz.rds")

Geospatial data wrangling

Combining Busstop and mpsz

Code chunk below populates the planning subzone code (i.e. SUBZONE_C) of mpsz sf data frame into busstop sf data frame.

busstop_mpsz <- st_intersection(busstop, mpsz) %>%
  select(BUS_STOP_N, SUBZONE_C) %>%
  st_drop_geometry()

Warning: attribute variables are assumed to be spatially constant throughout
all geometries

• st_intersection() is used to perform point and polygon overly and the output will be in point sf object.

• select() of dplyr package is then use to retain only BUS_STOP_N and SUBZONE_C in the busstop_mpsz sf data frame.

• five bus stops are excluded in the resultant data frame because they are outside of Singapore boundary.

datatable(busstop_mpsz)

Before we continue, we save the output in rds format:

write_rds(busstop_mpsz, "data/rds/busstop_mpsz.rds")  

Next, we are going to append the planning subzone code from busstop_mpsz data frame onto odbus6_9 data frame.

od_data <- left_join(odbus6_9 , busstop_mpsz,
            by = c("ORIGIN_PT_CODE" = "BUS_STOP_N")) %>%
  rename(ORIGIN_BS = ORIGIN_PT_CODE,
         ORIGIN_SZ = SUBZONE_C,
         DESTIN_BS = DESTINATION_PT_CODE)

Warning in left_join(odbus6_9, busstop_mpsz, by = c(ORIGIN_PT_CODE = "BUS_STOP_N")): Detected an unexpected many-to-many relationship between `x` and `y`.
ℹ Row 25632 of `x` matches multiple rows in `y`.
ℹ Row 673 of `y` matches multiple rows in `x`.
ℹ If a many-to-many relationship is expected, set `relationship =
  "many-to-many"` to silence this warning.

Before continue, it is a good practice for us to check for duplicate records.

duplicate <- od_data %>%
  group_by_all() %>%
  filter(n()>1) %>%
  ungroup()

If duplicated records are found, the code chunk below will be used to retain the unique records.

od_data <- unique(od_data)

It will be a good practice to confirm if the duplicating records issue has been addressed fully.

Next, we will update od_data data frame with the planning subzone codes.

od_data <- left_join(od_data , busstop_mpsz,
            by = c("DESTIN_BS" = "BUS_STOP_N")) 

Warning in left_join(od_data, busstop_mpsz, by = c(DESTIN_BS = "BUS_STOP_N")): Detected an unexpected many-to-many relationship between `x` and `y`.
ℹ Row 167 of `x` matches multiple rows in `y`.
ℹ Row 672 of `y` matches multiple rows in `x`.
ℹ If a many-to-many relationship is expected, set `relationship =
  "many-to-many"` to silence this warning.

duplicate <- od_data %>%
  group_by_all() %>%
  filter(n()>1) %>%
  ungroup()

od_data <- unique(od_data)

od_data <- od_data %>%
  rename(DESTIN_SZ = SUBZONE_C) %>%
  drop_na() %>%
  group_by(ORIGIN_SZ, DESTIN_SZ) %>%
  summarise(MORNING_PEAK = sum(TRIPS))

`summarise()` has grouped output by 'ORIGIN_SZ'. You can override using the
`.groups` argument.

We then save the output into rds file format:

write_rds(od_data, "data/rds/od_data.rds")

od_data <- read_rds("data/rds/od_data.rds")

Visualising Spatial Interaction

In this section, you will learn how to prepare a desire line by using stplanr package.

Removing intra-zonal flows

We will not plot the intra-zonal flows. The code chunk below will be used to remove intra-zonal flows.

od_data1 <- od_data[od_data$ORIGIN_SZ!=od_data$DESTIN_SZ,]

Creating desire lines

In this code chunk below, od2line() of stplanr package is used to create the desire lines.

flowLine <- od2line(flow = od_data1, 
                    zones = mpsz,
                    zone_code = "SUBZONE_C")

Creating centroids representing desire line start and end points.

Visualising the desire lines

To visualise the resulting desire lines, the code chunk below is used.

tm_shape(mpsz) +
  tm_polygons() +
flowLine %>%  
tm_shape() +
  tm_lines(lwd = "MORNING_PEAK",
           style = "quantile",
           scale = c(0.1, 1, 3, 5, 7, 10),
           n = 6,
           alpha = 0.3)

Warning in g$scale * (w_legend/maxW): longer object length is not a multiple of
shorter object length

Warning in g$scale * (x/maxW): longer object length is not a multiple of
shorter object length

When the flow data are very messy and highly skewed like the one shown above, it is wiser to focus on selected flows, for example flow greater than or equal to 5000 as shown below.

tm_shape(mpsz) +
  tm_polygons() +
flowLine %>%  
  filter(MORNING_PEAK >= 5000) %>%
tm_shape() +
  tm_lines(lwd = "MORNING_PEAK",
           style = "quantile",
           scale = c(0.1, 1, 3, 5, 7, 10),
           n = 6,
           alpha = 0.3)

Warning in g$scale * (w_legend/maxW): longer object length is not a multiple of
shorter object length

Warning in g$scale * (x/maxW): longer object length is not a multiple of
shorter object length